Acquisition of costly information in data-driven decision making

This paper formulates and solves an economic decision problem of the acquisition of costly information in data-driven decision making. The paper assumes an agent predicting a random variable utilizing several costly explanatory variables. Prior to the decision making, the agent learns about the relationship between the random variables utilizing its past realizations. During the decision making, the agent decides what costly variables to acquire and predicts using the acquired variables. The agent's utility consists of the correctness of the prediction and the costs of the acquired variables. To solve the decision problem, we split the decision process into two parts: acquisition of variables and prediction using the acquired variables. For the prediction, we propose an approach for training a single predictive model accepting any combination of acquired variables. For the acquisition, we propose two methods using supervised machine learning models: a backward estimation of the expected utility of each variable and a greedy acquisition of variables based on a myopic estimate of the expected utility. We evaluate the methods on two medical datasets. The results show that the methods acquire the costly variables efficiently.